Our lab has been focusing on making innovations in cutting-edge interdisciplinary fields for more than 20 years, and has sustained deep cultivation in computational optics, intelligent technology, optical computing, and brain science research. In recent years, to fill the gaps in mesoscale intravital imaging, we proposed a series of computational imaging methods. The RUSH series of microscopes developed based on this principle have achieved the world's first in vivo gigapixel fluorescence imaging. This technology has enabled breakthroughs in the observation of dynamic neuronal activities in the entire cortex of awake animals, supporting new discoveries in brain science and medicine, as well as providing new avenues for the development of artificial intelligence. Additionally, we proposed a new principle of scanning light field imaging and developed the framework of digital adaptive optics. We also put forward a series of spatiotemporal self-supervised learning methods, which have solved the century-old challenges such as phototoxicity, shot-noise limits, and optical aberrations in intravital imaging. These contributions enabled the first high-speed 3D observations of large-scale intercellular interactions at organ level in mammals across several hours, offering transformative tools for in vivo life science research. For data analysis, we have built up a new framework for multimodal and multiscale image analysis, such as the coupling between the spatial structures of complex organs and disease-related genetic mutations. To accelerate the processing of artificial intelligence, we proposed the first reconfigurable architecture of diffractive neural networks, developed an all-analog optoelectronic computing chip (ACCEL), and fabricated a large-scale interference-diffraction heterostructure integrated chip (TAICHI).